When Google rolls out a change it never announces, someone has to measure it from the outside. For years, that someone has often been Glenn Gabe, an independent search consultant based in Princeton, New Jersey. He runs G-Squared Interactive, also known as GSQi, and has been publicly documenting Google Search shifts since 2006.
Why our audience should care
Google confirms only a fraction of ranking changes and almost never tells a site why it lost visibility. For performance marketers and SEO teams, that creates a practical problem: you cannot separate a media planning failure from an organic visibility loss unless you know an algorithmic event actually happened. Gabe’s running record has become an early warning system for exactly that.
His publication rhythm is consistent. During the December 2025 broad core update, he posted daily Core Update Notes on X, built from visibility checks on more than 3,000 previously affected sites. Google announced that update on 11 December 2025; Gabe’s tracking indicated impact landed on 14 December, with a mid-rollout tremor on 20 December.
How he diagnoses a drop
After a traffic loss, Gabe starts with a delta report: a comparison of query and landing-page performance before and after the event, using Search Console API exports and third-party data from Semrush, Ahrefs and Sistrix. The goal is to isolate the most visible content that lost ground, because Google’s quality evaluation weighs what ranked rather than what simply exists.
His framework sorts drops into three causes that can look identical in a traffic graph:
- Relevancy adjustment: a page is judged less relevant for a query than before.
- Intent shift: Google decides a different type of result satisfies the query, which no site-side change can fix.
- Overall site quality: the only addressable category, and rarely a single defect.
On that third point, Gabe is direct. “There is never one smoking gun,” he says; the cause is usually multiple defects acting together. His checklist extends beyond content into advertising aggression, UX barriers and indexation hygiene. He also points to Navboost, the user interaction system described in US antitrust proceedings, as giving Google’s ranking systems a long memory for visitor satisfaction.
Put it to work
Treat unconfirmed visibility swings as signal, not verdict. When a drop appears, check whether Google’s status dashboard confirms a rollout, compare affected queries by intent, and look for multiple quality issues rather than one fix. If the change matches a documented update, hold off on reactive redesigns until the rollout finishes.
From Phantom to AI Mode
Gabe named the original Phantom update in May 2013 after sudden traffic collapses that Google would not acknowledge. The label stuck with Phantom 2 on 29 April 2015; Google denied an update at first, then confirmed a core ranking change affecting quality assessment. Search Engine Land renamed it the Quality Update, and Gabe later wrote that Google never adopted his name and does not care for it.
More recently, his work has moved toward enforcement tracking and AI search. He documented sponsored stores and quick web results inside AI Mode on 25 March 2026. On 28 April 2026, he flagged that an AdSense vignette trigger fires on browser back navigation, the exact behaviour Google’s spam policies had called back button hijacking two weeks earlier. And on 5 August 2026, he amplified reports of mass Blogger removals hours before Google acknowledged a bug.
Limitations marketers should know
Gabe’s method is correlational. Third-party visibility scores measure ranking across sampled keywords, not revenue, and sites that approach a recovery consultant are often sites that already lost traffic. Public update analysis also functions as an acquisition channel for a business selling recovery work, a commercial arrangement common across the discipline and rarely stated.
Recovery is not guaranteed by diagnosis. He documented improvement across roughly 400 sites damaged by the September 2023 Helpful Content Update when the June 2025 core update ran, nearly two years later. The widely cited full recovery of HouseFresh was called exceptional by its own editor against a population of publishers that did not survive.
Source: PPC Land
